« Back to Results
Marriott Marquis, Grand Ballroom 1
Hosted By:
American Economic Association
The second approach relies on a difference-in-differences design where the rate unionization at the state-industry level is used as a proxy for union threat effects. We estimate the impact of the rate of unionization on non-union wages using a distribution regression approach where the whole wage distribution is modelled as a function of a rich set of covariates that includes the rate of unionization as well as state, industry, year and state and industry trends. In this setting, unions threat/spillover effects are identified using state-industry specific trends in unionization rates and wages. We also consider an alternative estimation strategy where the rate of success of union organizing elections is used as a proxy for the threat effect. Estimates suggests that the threat of unionization tends to compress the non-union wage distribution. Using our preferred specification where the union threat is proxied using the unionization rate, we conclude that declining union threat/spillover effects account for close to 20% of the 1979-2017 growth in male wage inequality measured using the 90-50 gap or the standard deviation of log wages.
Labor Unions in the United States
Paper Session
Sunday, Jan. 5, 2020 10:15 AM - 12:15 PM (PDT)
- Chair: John Ahlquist, University of California-San Diego
The Effects of Union Membership on Political Participation: Evidence from Michigan Teachers Unions after Right to Work
Abstract
Labor union members turn out to vote at higher rates than non-members and are more likely to vote for Democrats. Does membership in a union cause members to be more politically engaged and more politically liberal? And what role do unions play in turning out their members and influencing their political views? To answer these questions, we study teachers unions in Michigan, exploiting changes in the legal landscape that weakened public sector unions. After Right to Work passed in Michigan in 2012, unions were no longer able to collect agency fees from their members. Dues-paying membership in the Michigan Education Association (MEA)---historically Michigan's second largest union---shrunk by 25 percent following RTW and political spending by the union was cut in half between 2012 and 2016. To understand the political effects of unions on the members themselves, we combine individual-level data on Michigan teachers, linking administrative data on educator certification to the Michigan voter file and campaign contribution data. Our data enable us to estimate individual-level event studies, comparing teacher party registration, voting, and contribution behavior before and after RTW, as well as direct effects of RTW on union membership in the MEA or other teachers unions. To sharpen our causal identification and to separate the messaging effects of RTW's passage from the actual implementation of RTW in local contracts, we exploit the staggered expiration of teachers collective bargaining agreements. Similar teachers working in different local school districts were exposed to the no-agency-fee provision of RTW at different times determined by when old contracts ended. We observe political participation in national, state, and local elections. Linking teachers to their family members and neighbors in the voter file, we are also able to estimate spillovers of union membership.Union Spillover Effects and Changes in Wage Inequality
Abstract
This paper looks at the contribution of union spillover effects in the growth in wage inequality over the 1979-2017 period. We consider several estimation strategies to estimate union spillover effects. We first use an event study approach that takes advantage of the introduction of right-to-work laws in three large Midwestern states (Indiana, Michigan, and Wisconsin) since 2011. The idea is that right-to-work laws weaken the power of unions and reduce the threat of unionization, which in turn may have an adverse impact on the wages of non-union workers. We find some evidence that the passage of right-to-work laws had an adverse impact on non-union wages, though it is hard to reach firm conclusions because of imprecise estimates.The second approach relies on a difference-in-differences design where the rate unionization at the state-industry level is used as a proxy for union threat effects. We estimate the impact of the rate of unionization on non-union wages using a distribution regression approach where the whole wage distribution is modelled as a function of a rich set of covariates that includes the rate of unionization as well as state, industry, year and state and industry trends. In this setting, unions threat/spillover effects are identified using state-industry specific trends in unionization rates and wages. We also consider an alternative estimation strategy where the rate of success of union organizing elections is used as a proxy for the threat effect. Estimates suggests that the threat of unionization tends to compress the non-union wage distribution. Using our preferred specification where the union threat is proxied using the unionization rate, we conclude that declining union threat/spillover effects account for close to 20% of the 1979-2017 growth in male wage inequality measured using the 90-50 gap or the standard deviation of log wages.
United States Intergenerational Mobility over the 20th Century: Evidence from Small Data
Abstract
A large body of evidence shows that labor unions reduce income inequality. Recent work in public economics, however, has highlighted that inequality in opportunity is as important as inequality in current income. We ask whether labor unions reduce inequality in opportunity – in addition to income – by estimating their effects on intergenerational mobility. We use historical survey data to construct new estimates of intergenerational mobility over the 20th century, taking advantage of all surveys that asked respondents about father's income. The use of survey data allows us to estimate intergenerational mobility in a consistent way over a long period, avoiding problems in historical record linkage as well as look at the mobility of women as well as men. The presence of a rich set of covariates unavailable in most administrative datasets also lets us examine which institutions mitigate intergenerational persistence across households. We find a substantial effect of union membership in attenuating intergenerational persistence, the largest of the covariates we consider. Our results highlight the importance of labor unions for the patterns of intergenerational mobility that have received so much recent attention.JEL Classifications
- J5 - Labor-Management Relations, Trade Unions, and Collective Bargaining
- J3 - Wages, Compensation, and Labor Costs